30th Annual San Antonio Breast Cancer Symposium.

Introduction

Over 8000 healthcare professionals involved in the treatment of breast cancer attended the international San Antonio Breast Cancer Symposium (SABCS) from 13-16 December 2007. At the opening address, Kent Osbourne outlined the future direction of breast cancer care including the 'personalisation' of cancer treatment with the use of genetics and genetic profiling for prognosis and prediction, targeted therapy and molecular mechanisms of resistance. Over 4 days, many aspects of breast cancer were highlighted, from diagnosis, to therapeutic management and survivorship issues.

Chemotherapy in early breast cancer

The benefit of adjuvant chemotherapy in early breast cancer was outlined early on in the meeting by Richard Peto on behalf of the Early Breast Cancer Trialists' Collaborative Group (EBCTCG) in the first Plenary Lecture. The most recently updated (2005-2006) worldwide Oxford Overview of adjuvant chemotherapy continued to demonstrate the benefit of chemotherapy across most age groups with oestrogen receptor (ER)-negative and ER-positive tumours, with greatest benefit in women <50 years of age. As chemotherapy schedules have evolved, there have been increasing gains in terms of reduction in death rates at 10 years from the use of cyclophosphamide, methotrexate and 5-fluorouracil (CMF) compared with controls (no chemotherapy), of anthracyclines compared with CMF, and of taxanes compared with anthracyclines. There was no apparent survival gain for those women treated with high-dose chemotherapy. A meta-analysis of individual patient data from randomised controlled trials (RCTs) of adjuvant high-dose chemotherapy (HDC) also supported this finding. D. Berry (MD Anderson Cancer Center, TX, USA; Abstr. 11) analysed 15 RCTs involving 6194 patients who underwent HDC versus standard chemotherapy schedules. Relapse-free survival was modestly improved by HDC [hazard ratio (HR), 0.87; 95% confidence interval (CI), 0.81-0.94; P=0.0005] but this did not extrapolate to improved overall

survival (HR, 0.94; 95% CI, 0.87-1.02; P=0.21). No subset of patients appeared to benefit more. However, the heterogeneity of intensity of chemotherapy and type of chemotherapy given may have obscured any potential gain.

It has become increasingly clear that some women with ER-positive tumours have a sufficiently good prognosis not to warrant additional chemotherapy, even in the presence of node-positive disease. Two presentations examined the potential identification of these individuals. Kathy Albain presented data on behalf of the Breast Cancer Intergroup of North America (Abstr. 10) on an analysis of the Southwest Oncology Group (SWOG) 8814 trial patients using the 21-gene recurrence assay. The 21-gene recurrence score (RS) assay is prognostic for women with nodenegative, ER-positive breast cancer receiving 5 years of tamoxifen; a high RS value predicts a large benefit from chemotherapy in node-negative disease, with no improvement if the RS is low. The aim of this analysis was to show the potential use of RS in a node-positive, ER-positive population. The SWOG 8814 study was a Phase III RCT of postmenopausal, node-positive, ER-positive breast cancer patients randomly assigned to one of three treatment arms: tamoxifen 5years, cyclophosphamide, adriamycin, 5-fluorouracil (CAF) chemotherapy concurrently or CAF sequentially with tamoxifen. Superior disease-free and overall survival at 10 years was demonstrated in the latter arm. RT-PCR was carried out for 16 genes plus five reference genes on paraffin-embedded specimens and an RS was calculated and disease-free survival (DFS) and overall survival (OS) derived. Analysis was limited to 367 samples from the tamoxifen and sequential CAF plus tamoxifen arms. The 21-gene RS was found to be prognostic for DFS and OS in the tamoxifen arm, and accurately predicted subgroups of patients who gained no benefit over time from CAF (low RS, <18; P=0.97) compared to a subgroup who gained strong benefit (high RS >31; P=0.033). The benefit was more pronounced over the first 5 years. The RS remained predictive when covariates such as age, tumour grade and HER2 status were included (but not when ER score was included in the analysis).

The prognostic utility of the 21-gene assay was also compared with Adjuvant! Online [1] in hormone receptor-positive operable breast cancer and 0-3 positive axillary lymph nodes treated with adjuvant chemohormonal therapy in a presentation by Lori Goldstein on behalf of the Eastern Cooperative Oncology Group (ECOG) and North American Cancer Intergroup (TBCI) (Abstr. 63). This comparison of Oncotype DX RS with classical clinico-pathological features as well as treatment factors integrated by Adjuvant! Online (with outcomes adjusted to 5 years), was carried out on a sample population from the E2197 study. Using a case-control analysis of 465 patients (of whom 99 had recurrent disease), prognostic accuracy using the different methods was compared. RS was a highly significant predictor of recurrence in both node-negative and node-positive disease, and there was little difference in recurrence rates in patients with 1 versus 0 positive lymph nodes. The prognostic accuracy of Adjuvant! Online varied depending on whether central grading results (more accurate prognosis) or local grading results were used, whereas RS was the most accurate in predicting recurrence. When used together, RS was able to identify those likely to recur despite a low to intermediate risk according to Adjuvant! Online. This interesting exploratory analysis combines two methods of risk assessment in the same group of patients. However, there was a relatively short median follow-up of the study population, and all the patients were treated with chemotherapy (including low RS/low risk according to Adjuvant! Online).

There appears to be increasing evidence that anthracyclines, the mainstay of most modern adjuvant schedules, may not be necessary for all individuals. Denis Slamon, on behalf of the Cancer International Research Group (CIRG) (Abstr. 13) presented the role of anthracycline-based chemotherapy in adjuvant therapy. He discussed evidence from in vitro cell lines that anthracycline sensitivity in HER2-positive tumours is due to co-amplification of the topoisomerase IIa (topo IIa) gene. This was extrapolated to an analysis of tumour samples ('the test set') from the 0648 registration trial for trastuzumab, in which patients were randomly assigned to doxorubicin and cyclophosphamide (AC) [+ or -] trastuzumab. The addition of trastuzumab significantly improved progression-free survival and OS. Subanalysis by topo IIa co-amplification identified in the AC alone arm a distinct subset of patients who greatly benefited from anthracycline use (survival, 18.2 to 38.5 months; P=0.004). This finding was validated in a study in which one of the arms contained no anthracycline, the Breast Cancer International Research Group (BCIRG) 006 study, and HER2-positive patients were randomly assigned to AC followed by docetaxel (no trastuzumab), AC, docetaxel and trastuzumab for 1 year, or docetaxel and carboplatin concurrent with trastuzumab (TCH). Overall, 2990 of the 3990 patients who took part were involved in the analysis, and topo IIa co-amplification was observed in 35% of patients. At the second interim analysis, DFS significantly improved in those treated with a taxane. However, if DFS was analysed by the presence of topo IIa co-amplification, the survival curves began to converge, indicating continued benefit of treatment in the non-trastuzumab arm in the presence of an anthracycline. Dr Slamon postulated that the benefits of anthracyclines lie with this small group of HER2-positive, topo IIa co-amplified patients, and that the majority of HER2-negative patients do not benefit as much from an anthracycline.

The role of taxanes was underscored by Stephen Jones, who presented the extended follow-up of the US Oncology Adjuvant Trial 9733 (US Oncology Resource Inc., Houston, TX; Abstr. 12) in which patients were randomly assigned to either AC or TC (docetaxel/cyclophosphamide). The 7-year follow-up to this study demonstrated ongoing DFS (HR, 0.74; P= 0.003) and OS (HR, 0.69; P=0.03) in favour of the TC arm. Overall, 16% of the patients were >65 years (n=160), and TC appeared to be equally effective in older patients with comparable, if not better, tolerance than to the AC schedule.

Paul Ellis discussed the surprising lack of taxane superiority observed in of the UK-based TACT (Taxotere as Adjuvant Chemotherapy Trial) study (Guy's, King's and St Thomas' Hospitals, London, UK; Abstr. 68). In total, 4162 women were randomly assigned between a standard anthracycline schedule [5-fluorouracil, epirubicin and cyclophosphamide (FEC) or epirubicin followed by CMF (Epi-CMF)] and four cycles of FEC followed by four cycles of taxotere (FEC-T). At a median follow-up of 50 months, DFS and OS outcomes were equivalent in both arms. Higher levels of grade 3/4 toxicity were seen in the taxane arm. This outcome, although in contrast to some of the major studies looking at the question of adjuvant taxane benefit, remains of importance when deciding adjuvant schedules for patients.

Endocrine therapy in early breast cancer

The updates from major adjuvant endocrine trials were presented, in addition to evidence of potential benefits of treatment of associated toxicity. John Forbes presented the 100-month follow-up from the Arimidex, Tamoxifen Alone or in Combination (ATAC) trial comparing 5 years of tamoxifen with 5 years of anastrazole in postmenopausal women with ER-positive breast cancer (University of Newcastle, NWS, Australia; Abstr. 41). This update provided information after completion of the treatment period; prior data had shown that anastrazole was more effective in terms of DFS, with fewer side effects and better tolerance than tamoxifen. At the time of analysis, approximately 1000 women with a mean age of 72 years had 9 years of follow-up and 3000 women had 8 years of follow-up. At 5 years, there was an absolute advantage of 2.8% in favour of anastrazole in time to recurrence in hormone receptor-positive patients, increasing to 4.8% after 9 years (HR, 0.75; P=0.01). This is the first demonstration of a carry-over effect with aromatase inhibitors (AIs). The estimated carry-over for AIs appears to be larger than that for tamoxifen (50% in years 5-9 as opposed to 33% for tamoxifen as estimated by the EBCTCG). Anastrazole also showed a significant absolute advantage of 2.4% in time to distant recurrence, and significant absolute difference of reduction in contralateral breast cancer of 1.7% (HR, 0.60; P=0.004). It is reassuring that the increased incidence of fractures observed with anastrazole during treatment appears to return to levels comparable to those in the tamoxifen arm after completion of treatment, with no relative increase in new mortality or morbidity.

Richard Peto presented preliminary results of the Adjuvant Tamoxifen: Longer Against Shorter (ATLAS) trial (University of Oxford, UK; Abstr. 48), in which 5 years versus 10 years of adjuvant tamoxifen was compared. This large (11,500 patients) multicentre study with 55% confirmed ER-positive patients, had reasonably complete data for 6-10 years and very limited data for 11-15 years. Nevertheless, the beginnings of a trend towards decreased recurrence and breast cancer mortality at 5+ years in the 10-year tamoxifen arm [ratio of rates with continued versus stopped tamoxifen, 0.79 (standard error, 0.13)] was observed, although no toxicity data or other details were available. Whether the final outcome of this study will still be relevant to today's practice of AI switching is yet unknown.

AI-induced oestrogen ablation accelerates bone loss and augments fracture risk in postmenopausal women, and several studies were reported at SABCS regarding early prevention with the use of bisphosphonates. The Zometa-Femara Adjuvant Synergy Trial (Z-FAST) evaluated the 36-month safety and efficacy of upfront versus delayed intravenous zoledronic acid (ZA, 4 mg every 6 months) in decreasing AI-associated bone loss in 602 postmenopausal women with early breast cancer (A Brufsky, Novartis Pharmaceuticals, NJ, USA; Abstr. 27). Patients in the delayed arm received ZA when bone density (BMD) measurements decreased to T-score -2 or clinical fracture occurred. All patients received calcium and vitamin D. Lumbar spine and total hip BMD increased for patients on upfront bisphosphonates, but decreased for patients on delayed treatment. By 36 months, 21% in the delayed arm had started ZA; however, 36-month fracture rates were not significantly different at 5.7% and 6.3% for the upfront and delayed arms, respectively. A substudy of the International Breast cancer Intervention Study (IBIS)-II prevention study was presented (S Singh, Wolfson Institute of Preventive Medicine, London, UK; Abstr. 28) in which postmenopausal women at high risk of breast cancer are randomly assigned to receive either anastrozole or placebo. Patients are then stratified according to T-score: those with normal T-score are observed; those with osteopenic scores between -2.5 and -1 are randomly assigned to either risedronate or placebo; and osteoporotic patients receive risedronate. In the absence of risedronate there is a fall in BMD at 1 year with an AI, which is not observed in osteopenic patients treated with an AI and residronate. A novel approach to bone loss on AIs was presented (G Ellis, Seattle Cancer Group Alliance, WA, USA; Abstr. 47) with the use of denosumab, a fully humanised highly specific monoclonal antibody that binds receptor activator of nuclear factor ?B (RANK) ligand. RANK ligand stimulates osteoclasts and bone reabsorption, and denosumab may reverse the AI-induced bone loss. In this Phase III, randomised, double-blind, placebo-controlled trial, 252 women on adjuvant AI therapy were randomly assigned to denosumab (60 mg subcutaneous every 6 months) or placebo. At 12 months, denosumab significantly improved lumbar spine BMD from baseline compared to placebo (P<0.0001). There was no difference between the study arms in terms of serious adverse events. By contrast, a study from the Austrian Breast and Colorectal Cancer Study Group (ABCSG) examined the role of bisphosphonates in premenopausal women treated with surgery and goserelin, and then randomly assigned to one of four treatment arms: tamoxifen, tamoxifen and ZA, anastrazole, or anastrazole and ZA. Again ZA was administered at 4 mg every 6 months. At 3 years and at 5 years (2 years after completion of treatment), patients who did not receive ZA experienced an average decrease in BMD of 11.3% and 6.8%, respectively, compared to baseline, whereas those who received ZA showed no change or an increase of 3.9% at the same time intervals. There were no differences between the arms in serious adverse events (M Gnant, on behalf of the ABCSG, Austria; Abstr. 26).

Radiotherapy

An authoritative overview of radiation advance and future directions in breast cancer was presented by Lori Pierce from the University of Michigan, as the second Plenary Lecture. Despite the benefits of adjuvant radiation therapy, 20-30% of women do not receive it after breast-conserving therapy due to cost or inconvenience. Typical standard whole-breast irradiation consists of 50 Gy in 25 daily fractions over a 5-week period. By increasing the fraction size and reducing the number of doses (and reducing the total delivered dose), accelerated hypofractionated whole-breast irradiation (AHWBI) may be simpler to deliver to more people after breast conservation. Timothy Whelan presented the 12-year median follow-up results of the Ontario Clinical Oncology Group (Juravinski Cancer Center, ON, Canada; Abstr. 21), in which patients were randomly assigned to AHWBI (42.5 Gy/16 fractions/22days) or the standard schedule. A total of 1234 patients were involved in the study, and investigators found that both arms of the study were statistically identical in terms of local recurrence and OS. Cosmetic outcome and Radiation Therapy Oncology Group (RTOG)/European Organization for Research and Treatment of Cancer (EORTC) late radiation morbidity were also equivalent. These results, along with the UK-based Standardisation of Radiotherapy (START) study, should encourage the more widespread use of AHWBI in breast cancer.

Sentinel lymph nodes

As sentinel lymph-node biopsy (SLNB) becomes the standard of care in most patients with early-stage breast cancer, there remains the question of whether all patients with positive SLNB require completion axillary lymph-node dissection (ALND). Two oral presentations approached this question from different perspectives. The National Surgical Adjuvant Breast and Bowel Project (NSABP)-32 study was a prospective study in which 5611 patients with clinically node-negative disease were randomly assigned to (1) SLNB followed by ALND or (2) SLNB followed by ALND only if there was a positive SLNB. A substudy analysis of the NSABP-32 trial by Thomas Julian (NSABP Medical Affairs, Biostatistical Centers, PA, USA; Abstr. 51) assessed the risk of non-sentinel lymph node positivity based on clinicopathological features of the tumour and findings from the positive SLNB. The number of positive sentinel lymph nodes, lymphovascular invasion, number of sentinel lymph nodes examined, clinical tumour size, number of hotspots and treatment type were all significant predictors of positive non-sentinel lymph nodes, but age, type of biopsy, histological grade, HER2 status, tumour location and hormone receptor status were not. Even patients with favourable markers still had a 10% chance of having a positive non-sentinel lymph node. Nevertheless, the investigators concluded that there are a group of patients with very small tumours (<1 cm), no lymphovascular invasion and only a single positive sentinel lymph node out of multiple sentinel nodes (>5) who could potentially avoid an ALND. However, no data on size of metastases and presence of micrometastases in the node-negative SLNBs were available.

The American Joint Committee on Cancer (AJCC) Cancer Manual (6th edition) delineates axillary metastases by size: macrometastases >2 mm; micrometastases >0.2 to 2 mm; and isolated tumour (ITC) cells <0.2 mm. Nora Hansen (John Wayne Cancer Institute, CA, USA; Abstr. 52) presented the results of a study in which 790 patients were categorised by the size of the nodal metastases (negative, ITC, micro and macro-metastases). There was no statistically significant difference in DFS and OS among patients who had nodal disease size less than 2 mm, and all these patients had significantly better outcomes than patients with macrometastases. However, 25% of node-negative patients received adjuvant chemotherapy and therefore the true outcomes for these pathologically smaller nodal burdens may not be apparent from such a study.

Other highlights

Exciting preclinical results were presented by Sophia Ran using nab-paclitaxel and/or bevacizumab to treat lymphatic and pulmonary metastases of human breast cancer cells in female severe combined immunodeficiency (SCID) mice (Southern Illinois University, IL, USA; Abstr. 74). Cells that survive chemotherapy have upregulated vascular endothelial growth factor (VEGF)-A, and may potentiate tumour recurrence and spread. Treatment with an anti-VEGF-A (bevacizumab) and a cytotoxic agent (nab-paclitaxel) potentially overcomes this. Nab-paclitaxel is a nanoparticle form of paclitaxel that is coated with albumin and has reduced toxicity and enhanced drug delivery. For this study, MDA-MB-231 human breast cancer cells were orthotopically implanted into mammary fat pads of mice and the tumours allowed to grow, with lymphatic and pulmonary metastases developing also. Although bevacizumab and nab-paclitaxel when given separately resulted in some responses in the primary tumour and metastases, synergistic effects were observed when both drugs were given, resulting in complete eradication of the primary tumours and metastases.

Pharmacogenetics examines the inherited variants in DNA that predispose patients to adverse drug reactions. Drug efficacy and safety may be affected by a single nucleotide polymorphism (SNP) in an area coding for drug target, transporter or drug metabolising enzyme. Cytochrome P4502D6 has been a recent focus of attention due to its important role in metabolising tamoxifen to its active metabolite endoxifen. Several genetic variations of 2D6 have been observed, and endoxifen levels correlate with the variants. Drugs such as fluoxetine and paroxetine are known to inhibit 2D6. Further data on the 2D6 story was presented by James Rae (on behalf of the COBRA investigators; Abstr. 77) in which compliance to tamoxifen treatment is influenced by 2D6 activity. A total of 297 patients from the Consortium on Breast Cancer Pharmacogenomics (COBRA) tamoxifen trial were genotyped for CYP450. Patients who most successfully metabolised tamoxifen to endoxifen were more likely to have tamoxifen-related side effects, and more likely to withdraw from the study, abrogating the tamoxifen benefits.

Conclusion

SABCS continues to provide a comprehensive arena for the presentation of research, as well as state-of-the-art management of breast cancer. In addition, mini-symposia by international opinion leaders were given on topics such as stem cells, hormone-receptor therapy and breast cancer incidence, and targeting the HER2 network.

Access to much of the content is available on-line via the SABCS website [2]. Abstracts in full and selected audio and video highlights of SABCS are again freely accessible. In addition this year, an oral presentation slide reviewer and Poster2view sections have been included.

All the abstracts appear in an expanded issue of Breast Cancer Research and Treatment, 2007, 106 (suppl 1).

References

[1.] Adjuvant! Online. www.adjuvantonline.com (accessed 14 February 2008).

[2.] San Antonio Breast Cancer Symposium. www.sabcs.org (accessed 14 February 2008).

Marina Parton

Royal Marsden Hospital and Kingston NHS Foundation Trusts, UK

Correspondence to: Marina Parton Royal Marsden Hospital, Fulham Road, London SW3 6JJ, UK (email: [e-mail])